Where's MIMO?

Political standoff in IEEE slows progress of much-anticipated 100M bit/sec WLAN standard.

Political standoff in IEEE slows progress of the much-anticipated 100M bit/sec WLAN standard called MIMO.

There is one great truth in anything high tech -- faster is always better. Which is why customers are eagerly anticipating 802.11n wireless LAN products that promise greater throughput, not to mention greater range and reliability, than today's 801.11a/b/g/ products.

But if you're looking to buy standards-based, Wi-Fi Alliance-approved, enterprise-ready 802.11n Multiple-Input, Multiple-Output (MIMO) gear, you'll will have to wait while competing vendor groups hash out their differences.

Even though the 802.11n effort has been ongoing within the IEEE since 2002, don't expect to see a final standard nailed down until the end of 2006 and don't expect to see products until 2007. For example, 802.11n handsets for voice over Wi-Fi won't be available for at least a year after the standard is finished, because of the complexity of engineering such a device.

The IEEE standards development process is designed to assure that standards are broadly acceptable and have sufficient validity to serve even in legal proceedings.

The process is rigorous from the start. Before the IEEE approves a standards effort, a proposal must meet several criteria: broad market potential, inter-layer compatibility, an identity distinct from other 802 standards, technical feasibility and economic feasibility.

Once a Program Authorization Request (PAR) is approved, subsequent work is subject to five principles: due process, consensus (it takes a 75% vote to pass a standard), openness, balance (broad representation of interests) and the right of appeal.

The PAR for MIMO was approved in September 2003. The group initially received 36 proposals, which were reduced to four in January 2005, then to two.

TGnSync vs. WWiSE (vs. EWC?)

One proposal called TGnSync includes Agere, Atheros, Cisco, Intel, Qualcomm and Symbol Technologies. The key TGnSync position is the support of wider bandwidth channels (40 MHz vs. 20 MHz used in 802.11), potentially simplifying the design of standards-based products.

The other proposal is called WWiSE and includes Airgo Networks (the first company to build a pre-802.11n or MIMO Enhanced WLAN [MEW] chipset), Broadcom, Conexant, HP, Motorola, Siemens and Texas Instruments. An interesting element of WWiSE is a royalty-free contribution of intellectual property, potentially lowering costs for products based on the standard. Both groups were asked to meet offline and work out their differences, with a final agreement expected by next month, because neither group had been able to get 75% approval.

However, a new vendor group, called the Enhanced Wireless Consortium, unexpectedly emerged within the past couple of weeks. Depending how it's viewed, this new group could throw a major monkey wrench into the process, or it could break the deadlock.

It appears that EWC, which plans to put its proposal before the IEEE in hopes of reaching the 75% threshold, includes companies from both the WWiSE and TGnSync camps. At the chip level, the new group includes Intel, Atheros, Broadcom and Marvell, but not Airgo.

In the best-case scenario, a single proposal could be ready by next month and approved as a draft standard in January. That would mean three years between PAR approval and a draft standard. By contrast, 802.11b took only two years.

But the availability of a draft standard is not the same as completion of a standard. Much additional work is needed, including ferreting out errors and inconsistencies.

There also are letter ballots (within the .11n Task Group), Sponsor Ballots (think of this as external peer review), a vote by the 802 Executive Board and a final vote by the IEEE Standards Board. Once all that is done, it takes a few weeks for the final standard to be published.

If all goes well,we're talking the end of 2006.

Spec approval doesn't mean that approved products are available. The Wi-Fi Alliance plays a role almost as important as that of the IEEE. Because the standards do not include specifications for test procedures (or the tests themselves) to verify compliance, compatibility or interoperability, the Wi-Fi Alliance conducts interoperability testing and certifies products. While its work can proceed in parallel with overall standards development, this effort is separate from 802.11.

With standards-based MIMO on the horizon and with non-standard products out now, what should enterprise executives do?

First of all, standards-based 802.11n products will be backward-compatible with existing 802.11/Wi-Fi products, so we recommend continuing to purchase traditional products in advance of the availability of Wi-Fi approved 802.11n products. That's if you can wait for 100M bit/sec performance.

When it comes to the products on the market today based on MEW (formerly "pre-n") technologies, the issue is a bit trickier.

We recommend enterprise use of MEW clients today, but the lack of any enterprise-specific MEW-based infrastructure components gives us pause. At a minimum, enterprise-class access points must support power over Ethernet and management capabilities, and ideally be part of a switched or centralized WLAN architecture. We don't expect to see many of these in advance of the standard, but availability of the draft standard will spur the production of prestandard products.

This being the case, we don't put too much stock in claims of upgradeability to the standard for MEW products. While firmware upgrades are likely, full compliance with the final, Wi-Fi-certified standard cannot be guaranteed.

While MEW products should have a happy life ahead of them for the next 18 months or so, even in some enterprise client applications, it's very likely that the enterprise will need to plan a phased transition to fully compliant 802.11n products that could take several years.

One other important point - standards are specifications for interfaces, but not recipes for how to design and build successful products. We expect to see a broad range of products, some clearly better than others, and some designed for specific applications. We expect a high degree of variability in 802.11n products, including some offering well in excess of 300M bit/sec, and perhaps as much as 600M bit/sec.

And this begs the question: Is 802.11n the end of the road for WLANs? After all, 600M bit/sec should be plenty for a huge number of applications. Still, innovation continues to define wireless as a whole. Ultrawideband technologies promise multiple-gigabit throughput, and 60 GHz radios could offer more than that.

The core justification for the adoption of wireless has always been convenience. The anytime/anywhere nature of wireless, and the lack of any substitute technology, will continue to guarantee its success in the future. WLANs, largely because of 802.11 and the work of the Wi-Fi Alliance, have continued their rapid advance in pursuit of the goal of parity - both in throughput and overall functionality -with wire.

With 802.11n on the horizon, it's hard to imagine that WLANs will not become the default connectivity for the vast majority of enterprise users over the next few years.

Mathias is a principal at the Farpoint Group. He can be reached at craig@farpointgroup.com.

Timeline on MIMO effort

Learn more about this topic

New MIMO Wi-Fi to outpace Ethernet

09/14/05

IEEE group scraps 802.11n proposalMIMO products boost 802.11g nets

05/19/05

03/21/05

Stage set for compromise on IEEE high-speed wireless

03/21/05

Join the discussion
Be the first to comment on this article. Our Commenting Policies